Assorted links

Levine is exactly right. There is so much confusion over behavioral econ. The rational hypothesis is not a hypothesis of psychology. Deviations from rationality are far fewer than the pop lot suggests.

Which is not to say that neoclassical assumptions can’t be improved. On my view at least, the low-hanging fruit is the assumption of ergodicity and uniformity inherent in how the rational hypothesis is operationalized through revealed preferences. Bubbles and crashes are not irrational! They are the consequence of low-frequency, large-scale effects gradually dominating short-term, local dynamics.

The way forward for economics is to think bigger and longer term. Psychology is wonderful and important. And economists can learn much from pyschologists. But economics is not psychology.

There’s lots of good things in this Levine article. For one, I think it’s absolutely true that behavioral economics has been giving a free pass in the area of laying the same kind of scrutiny against it that it has against neoclassical assumptions. Maybe that is changing. I liked his discussion, too, of impulsive behavior versus rational behavior, and how in fact they are not in conflict. This quote, for instance, was very interesting: “The point here is simple: our “rational† self is not intrinsically in conflict with our impulsive self. In fact the evidence is that our rational self often facilitates rather than overrides the activities of our impulsive self.”

Levine scores a few points, but is very far from being “exactly right,”
not that behavioral econ should not be subjected to the sorts of analysis
that we should also subject conventional economic theory to as well. But,
there are many distortions and misrepresentations in the Levine article.

So, “rationality” is defined as Nash equilibrium behavior. But we have
seen many repeated prisoners’ dilemma experiments where participants
cooperate in contradiction to the Nash equilibrium of cheating. This old
bugaboo that upset John Nash himself when he saw it happen back at Rand
in the early 50s with the original Dresher-Flood experiments is not
mentioned by Levine. Of course, going along with this is that in many
situations there are multiple Nash equilibria (and almost anything can
be “explained” by quantal response equilibria, so whoop-de-doo!). This
certainly complicates disproving “rationality” in general.

QRE is dragged in to explain ultimatum game behavior. Somehow, “spite”
becomes “rational,” which it may be in some long-term evolutionary sense
as argued by many. But it is not short-run Nash behavior, not at all,
and the average offer of 40% in UGs is not easily explained by QRE,
although in the end Levine invokes “learning” without showing it solves
anything, although there is evidence that offers decline over time in
repeated UGS, even as some participants do not lower their offers.

Yes, there are some situations where conventional theory works fine,
even arguably better than itself. One is the old rapid convergence to
competitive equilibrium in double auctions. Fine. Another is the voting
experiment, and there are situations where errors by befuddled individuals
offset each other in law of large numbers ways to give group rationality
in spite of individual irrationality. But this does not always work, as
the failure of the linear bivariate Gaussian copula in the recent CDS
markets proves, which was based on assuming such offsetting of errors in
the housing mortgage markets (or lack of positive correlation across them).

Plott and Zeiler are invoked to dismiss the enormous evidence on WTA-WTP
gap. I will not cite specifics, but I have seen papers now forthcoming
in journals that critique P and Z. The strongest evidence for WTA-WTP
gap is from mountains of field data dating back to the 1970s, much of it
involving contingent valuation studies of environmental public goods. The
critique by P and Z of some of the framing in some studies of students
giving each other coffee cups does not begin to undo that mountain of data.

We have this bizarre effort to reconcile “impulsive” behavior with
rationality. The method is to cite that Spitzer and Limbaugh actually
planned their “impulsive” behavior ahead of time. So what? Then their
behavior was not impulsive, but that of many others is. Indeed, one of
the more ubiquitous findings of behavioral econ is that of hyperbolic
discounting, which can be invoked to “explain” impulsive behavior. Let us
be clear: hyperbolic discounting is inconsistent with time consistency,
which one must have for the stronger usual forms of “rationality.”

However, there might be another way of reconciling “rationality” with
impulsivity, which would be redefine it away from simple Nash equilibrium
into some sort of long run evolutionary fitness. So, from neuroeconomics,
which Levine seems to diss, we know that different (and lower down) parts
of the brain get fired when we engage in impulsivity (I want that food,
sex, pleasure, escape from danger NOW!). Sometimes such impulsivity pays
off big time (get to reproduce with the partner of my dreams by coming on
passionately and spontaneously!), even as sometimes it does not work out
so well under “normal” circumstances.

Oh yes, and there are a number of other alternatives to expected utility
than prospect theory. Frankly, this is a piss poor paper not worthy of
publication in even a fourth rate journal, although given the prominence
of the author, it will probably score better than that.

The “TARP wife” thing is a pretty tired Wall St in-joke. Any decent economist should be able to figure out that no one on Wall St is avoiding good restaurants to stay below the sumptuary radar or to make rent.

Per Se was quite nice last time I went. Most people there seemed time, not money, constrained.